Search

Abstract Deep learning models demonstrate impressive performance in rapidly predicting urban floods, but there are still limitations in enhancing physical connectivity and interpretability. This study proposed an innovative modeling approach that integrates convolutional neural networks with weighted cellular automaton (CNN-WCA) to achieve the precise and rapid prediction of urban pluvial flooding processes and enhance the physical connectivity and reliability of modeling results. The study began by generating a rainfall-inundation dataset using WCA and LISFLOOD-FP, and the CNN-WCA model was trained using outputs from LISFLOOD-FP and WCA. Subsequently, the pre-trained model was applied to simulate the flood caused by the 20 July 2021 rainstorm in Zhengzhou City. The predicted inundation spatial distribution and depth by CNN-WCA closely aligned with those of LISFLOOD-FP, with the mean absolute error concentrated within 5 mm, and the prediction time of CNN-WCA was only 0.8% that of LISFLOOD-FP. The CNN-WCA model displays a strong capacity for accurately predicting changes in inundation depths within the study area and at susceptible points for urban flooding, with the Nash-Sutcliffe efficiency values of most flood-prone points exceeding 0.97. Furthermore, the physical connectivity of the inundation distribution predicted by CNN-WCA is better than that of the distribution obtained with a CNN. The CNN-WCA model with additional physical constraints exhibits a reduction of around 34% in instances of physical discontinuity compared to CNN. Our results prove that the CNN model with multiple physical constraints has significant potential to rapidly and accurately simulate urban flooding processes and improve the reliability of prediction.

ObjectivesWhen assessing the safety of the connectors of a multi-module offshore floating power generation platform, in order to compensate for the inability to carry out the real-time monitoring of the structural stress field across the whole domain due to a limited numbers of sensor, a digital twin method based on a simulation database is proposed that can rapidly predict the platform's stress field. MethodsBy downgrading the three-dimensional physical model of the connectors to a one-dimensional digital model, the stress field data is interpolated and deduced in digital space, thereby achieving the rapid prediction of the structural stress field across the whole domain and its visual display.ResultsThe results show that the simulation model is in good agreement with the test results, with a maximum absolute error of 8.61%; for the interpolation of data under different loading angles, when the interpolation step of the loading angle is 10°, the aver-age absolute error of stress is 1.98%; and for the interpolation of data under different loads, when the interpolation step of the load is 10 t, the average absolute error of stress is 1.28%, achieving the rapid prediction and visualization of the connectors' stress field distribution. Conclusions The digital twin-based model of connectors can provide useful references for the rapid dynamic perception and scientific prediction of the structural strength of offshore floating power generation platforms.

Deep learning models demonstrate impressive performance in rapidly predicting urban floods, but there are still limitations in enhancing physical connectivity and interpretability. This study proposed an innovative modeling approach that integrates convolutional neural networks with weighted cellular automaton (CNN-WCA) to achieve the precise and rapid prediction of urban pluvial flooding processes and enhance the physical connectivity and reliability of modeling results. The study began by generating a rainfall-inundation dataset using WCA and LISFLOOD-FP, and the CNN-WCA model was trained using outputs from LISFLOOD-FP and WCA. Subsequently, the pre-trained model was applied to simulate the flood caused by the 20 July 2021 rainstorm in Zhengzhou City. The predicted inundation spatial distribution and depth by CNN-WCA closely aligned with those of LISFLOOD-FP, with the mean absolute error concentrated within 5 mm, and the prediction time of CNN-WCA was only 0.8% that of LISFLOOD-FP. The CNN-WCA model displays a strong capacity for accurately predicting changes in inundation depths within the study area and at susceptible points for urban flooding, with the Nash-Sutcliffe efficiency values of most flood-prone points exceeding 0.97. Furthermore, the physical connectivity of the inundation distribution predicted by CNN-WCA is better than that of the distribution obtained with a CNN. The CNN-WCA model with additional physical constraints exhibits a reduction of around 34% in instances of physical discontinuity compared to CNN. Our results prove that the CNN model with multiple physical constraints has significant potential to rapidly and accurately simulate urban flooding processes and improve the reliability of prediction. [ABSTRACT FROM AUTHOR]

In the present study, a data-driven neural network methodology was proposed for the rapid prediction of multi-physics interactions in heat pipe reactors, and grounded in a comprehensive framework of multi-physics coupling. By leveraging neural networks, certain numerical computation modules within the framework were effectively replaced, which facilitated an efficient resolution of the iterative process inherent to the nuclear-thermal-mechanical multi-physics coupling. In the process of multi-physics coupling, there exists a sophisticated iterative mechanism, wherein the computational analyses of distinct physical domains largely function autonomously. When contemplating the integration of neural network paradigms as surrogates within the multi-physics coupling framework, it is pivotal to target the numerical computations intrinsic to individual physical domains. By deploying one or multiple neural network models as substitutes, this methodology not only sustains the scalability inherent to rapid predictive schemes but also attenuates the complexity associated with the neural network’s surrogate functionalities. Concurrently, this approach necessitates a reduced dataset for training, and through the amalgamation of diverse neural network models, markedly augments the adaptability of the overarching rapid predictive framework. Furthermore, multiple neural networks, operating autonomously, can be dynamically calibrated with respect to their training data volumes based on their predictive performance. This not only optimizes their predictive accuracy but also facilitates the cost-effective expansion of training datasets in accordance with specific model requirements. Taking “Megapower” as the subject of analysis, pertinent neural network models were constructed and employed for rapid predictive parameter estimation, with the results juxtaposed against traditional numerical computations. The discrepancies between the key parameters yielded by rapid predictions and those from numerical simulations were minimal. Specifically, the maximum stress differential does not exceed 2 MPa, and the average fuel temperature difference is less than 3 K. Notably, while numerical computations demand six hours of processing time, the rapid prediction on the same platform requires a mere 4 minutes, marking a reduction in computational time by over 95%. Given these outcomes, it can be posited that neural network-based rapid prediction schemes exhibit high precision and accelerated processing speeds. Coupled with their inherent flexibility and modest training data requirements, and the capacity for targeted model optimization, it is advocated that neural network-based rapid prediction serves as a preferred methodology for addressing challenges such as core optimization in scenarios that demand large-scale computations. Concurrently, the neural network models utilized in rapid predictions have less stringent computational resource requirements and offer greater adaptability to varied environments. Capitalizing on these attributes, in conjunction with contemporary compact mobile reactors, one can facilitate on-site deployments with relative ease. Such integrations pave the way for prompt alerts in accident scenarios, thereby enhancing the safety of the reactor core.

Urban floods pose a significant threat to human communities, making its prediction essential for comprehensive flood risk assessment and the formulation of effective resource allocation strategies. Data-driven deep learning approaches have gained traction in urban emergency flood prediction, addressing the efficiency constraints of physical models. However, the spatial structure of rainfall, which has a profound influence on urban flooding, is often overlooked in many deep learning investigations. In this study, we introduce a novel deep learning model known as CRU-Net equipped with an attention mechanism to predict inundation depths in urban terrains based on spatiotemporal rainfall patterns. This method utilizes eight topographic parameters related to the height of urban waterlogging, combined with spatial rainfall data as inputs to the model. Comparative evaluations between the developed CRU-Net and two other deep learning models, U-Net and ResU-Net, reveal that CRU-Net adeptly interprets the spatiotemporal traits of rainfall and accurately estimates flood depths, emphasizing deep inundation and flood-vulnerable regions. The model demonstrates exceptional accuracy, evidenced by a root mean square error of 0.054 m and a Nash–Sutcliffe efficiency of 0.975. CRU-Net also accurately predicts over 80% of inundation locations with depths exceeding 0.3 m. Remarkably, CRU-Net delivers predictions for 3 million grids in 2.9 s, showcasing its efficiency. HIGHLIGHTS The incorporation of spatial rainfall distribution into urban flood forecasting models is performed in this study.; Integration of attention mechanisms enhanced the identification of high-risk flood areas.; The developed model can predict rapid floods within seconds.;

Copyright of Atomic Energy Science & Technology is the property of Editorial Board of Atomic Energy Science & Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In this paper, a POD reduced-order interpolation model for solving the in situ pyrolysis process of tar-rich coal is employed to predict the flow and heat transfer performance in the porous media region so as to save computational resources and realize fast calculations. Numerical simulation using the finite volume method (FVM) is firstly used to obtain sample data, based on the samples through the primary function and spectral coefficients of the solutions. The physical field information and parameter distribution under different conditions of inlet temperature, inlet velocity and permeability are predicted. The results are compared with those of FVM to verify the accuracy of the calculated results. The relative mean deviation (RME) of the results of the POD prediction of each parameter for each working condition was synthesized to be no more than 5%. The performance of in situ pyrolysis of tar-rich coal is then investigated, and the oil and gas production are predicted. As the inlet velocity increases from 0.3 m/s to 0.9 m/s, the fraction of high-quality oil and gas production reaches 0.47 and then decreases to 0.38. Increasing the inlet temperature and permeability has a negative effect on the fraction of high-quality hydrocarbon production, after which the quality fraction of high-quality oil and gas dropped sharply to about 0.22. Porosity has a positive impact on the oil and gas production. When the porosity reaches 0.3, the quality fraction of high-quality oil and gas can reach 0.27. [ABSTRACT FROM AUTHOR]

Tea polyphenol and epigallocatechin gallate (EGCG) were considered as key components of tea. The rapid prediction of these two components can be beneficial for tea quality control and product development for tea producers, breeders and consumers. This study aimed to develop reliable models for tea polyphenols and EGCG content prediction during the breeding process using Fourier Transform–near infrared (FT-NIR) spectroscopy combined with machine learning algorithms. Various spectral preprocessing methods including Savitzky–Golay smoothing (SG), standard normal variate (SNV), vector normalization (VN), multiplicative scatter correction (MSC) and first derivative (FD) were applied to improve the quality of the collected spectra. Partial least squares regression (PLSR) and least squares support vector regression (LS-SVR) were introduced to establish models for tea polyphenol and EGCG content prediction based on different preprocessed spectral data. Variable selection algorithms, including competitive adaptive reweighted sampling (CARS) and random forest (RF), were further utilized to identify key spectral bands to improve the efficiency of the models. The results demonstrate that the optimal model for tea polyphenols calibration was the LS-SVR with Rp = 0.975 and RPD = 4.540 based on SG-smoothed full spectra. For EGCG detection, the best model was the LS-SVR with Rp = 0.936 and RPD = 2.841 using full original spectra as model inputs. The application of variable selection algorithms further improved the predictive performance of the models. The LS-SVR model for tea polyphenols prediction with Rp = 0.978 and RPD = 4.833 used 30 CARS-selected variables, while the LS-SVR model build on 27 RF-selected variables achieved the best predictive ability with Rp = 0.944 and RPD = 3.049, respectively, for EGCG prediction. The results demonstrate a potential of FT-NIR spectroscopy combined with machine learning for the rapid screening of genotypes with high tea polyphenol and EGCG content in tea leaves. [ABSTRACT FROM AUTHOR]

Abstract Climate change has led to increasing frequency of sudden extreme heavy rainfall events in cities, resulting in great disaster losses. Therefore, in emergency management, we need to be timely in predicting urban floods. Although the existing machine learning models can quickly predict the depth of stagnant water, these models only target single points and require large amounts of measured data, which are currently lacking. Although numerical models can accurately simulate and predict such events, it takes a long time to perform the associated calculations, especially two-dimensional large-scale calculations, which cannot meet the needs of emergency management. Therefore, this article proposes a method of coupling neural networks and numerical models that can simulate and identify areas at high risk from urban floods and quickly predict the depth of water accumulation in these areas. Taking a drainage area in Tianjin Municipality, China, as an example, the results show that the simulation accuracy of this method is high, the Nash coefficient is 0.876, and the calculation time is 20 seconds. This method can quickly and accurately simulate the depth of water accumulation in high-risk areas in cities and provide technical support for urban flood emergency management.

Climate change has led to increasing frequency of sudden extreme heavy rainfall events in cities, resulting in great disaster losses. Therefore, in emergency management, we need to be timely in predicting urban floods. Although the existing machine learning models can quickly predict the depth of stagnant water, these models only target single points and require large amounts of measured data, which are currently lacking. Although numerical models can accurately simulate and predict such events, it takes a long time to perform the associated calculations, especially two-dimensional large-scale calculations, which cannot meet the needs of emergency management. Therefore, this article proposes a method of coupling neural networks and numerical models that can simulate and identify areas at high risk from urban floods and quickly predict the depth of water accumulation in these areas. Taking a drainage area in Tianjin Municipality, China, as an example, the results show that the simulation accuracy of this method is high, the Nash coefficient is 0.876, and the calculation time is 20 seconds. This method can quickly and accurately simulate the depth of water accumulation in high-risk areas in cities and provide technical support for urban flood emergency management. [ABSTRACT FROM AUTHOR]

Total volatile basic nitrogen (TVB-N) level rapid evaluation on chicken meat based on gas sensor array (GSA) technique was studied in this paper. GSA responses to chicken meat stored at 4°C were examined for 5 days. TVB-N content was synchronously measured by chemical examination. Principal component analysis (PCA) and non-linear double-layered cascaded serial stochastic resonance (DCSSR) were utilized for measurement data analysis. TVB-N examination results suggested that chicken meat stored for more than 3 days was not fresh. PCA showed poor discrimination abilities, while DCSSR signal-to-noise ratio (SNR) quantitatively characterized the freshness of all samples. Chicken meat TVB-N forecasting model was developed by non-linear fitting between SNR eigenvalues and TVB-N values. The predicting model was constructed. Validation experiment results demonstrated that the forecasting accuracy of the developed model reached 93.3%.

During the prediction of the tool point frequency response function (FRF) based on receptance coupling substructure analysis (RCSA), in allusion to the problems that the existing tool model cannot simulate the contour of the tool precisely and the large amount of calculation for joint parameter identification during the calculation of the rotational FRF of the spindle, the joint model and the optimization algorithm affect the identification accuracy when acquiring the rotational FRF of the spindle. This study proposes a rapid prediction method for the tool point FRF on the basis of RCSA. The method divides the spindle-holder-tool system into two parts: the spindle-holder-partial shank and the remaining tool. On this basis, this study proposes a method of accurately establishing the model of the remaining tool with a 3D scanner and improves the calculation method of the rotational FRF of the spindle-holder-partial shank assembly through the reverse RCSA and finite difference method directly through the modal impact testing on the cutting tool. This study provides the experiments on a vertical machining center and compares the predicted FRF and the measured FRF on the tool point. The result indicates that the predicted FRF curve is consistent with the measured FRF curve with the relative error of natural frequency within 4.0%. This study confirms that the method is feasible, effective, and convenient. [ABSTRACT FROM AUTHOR]

Rice straw is a major kind of biomass that can be utilized as lignocellulosic materials and renewable energy. Rapid prediction of the lignocellulose (cellulose, hemicellulose, and lignin) and organic elements (carbon, hydrogen, nitrogen, and sulfur) of rice straw would help to decipher its growth mechanisms and thereby improve its sustainable usages. In this study, 364 rice straw samples featuring different rice subspecies (japonica and indica), growing seasons (early-, middle-, and late-season), and growing environments (irrigated and rainfed) were collected, the differences among which were examined by multivariate analysis of variance. Statistic results showed that the cellulose content exhibited significant differences among different growing seasons at a significant level (p <0.01), and the contents of cellulose and nitrogen had significant differences between different growing environments (p<0.01). Near infrared reflectance spectroscopy (NIRS) models for predicting the lignocellulosic and organic elements were developed based on two algorithms including partial least squares (PLS) and competitive adaptive reweighted sampling-partial least squares (CARS-PLS). Modeling results showed that most CARS-PLS models are of higher accuracy than the PLS models, possibly because the CARS-PLS models selected optimal combinations of wavenumbers, which might have enhanced the signal of chemical bonds and thereby improved the predictive efficiency. As a major contributor to the applications of rice straw, the nitrogen content was predicted precisely by the CARS-PLS model. Generally, the CARS-PLS models efficiently quantified the lignocellulose and organic elements of a wide variety of rice straw. The acceptable accuracy of the models allowed their practical applications. [ABSTRACT FROM AUTHOR]

The open circuit voltage of lithium ion batteries in equilibrium state, as a vital thermodynamic characteristic parameter, is extensively studied for battery state estimation and management. However, the time-consuming relaxation process, usually for several hours or more, seriously hinders the widespread application of open circuit voltage. In this paper, a novel voltage relaxation model is proposed to predict the final open circuit voltage when the lithium ion batteries are in equilibrium state with a small amount of sample data in the first few minutes, based on the concentration polarization theory. The Nernst equation is introduced to describe the evolution of relaxation voltage. The accuracy and effectiveness of the model are verified using experimental data on lithium ion batteries with different kinds of electrodes (LiCoO2/mesocarbon-microbead and LiFePO4/graphite) under different working conditions. The validation results show that the presented model can fit the experimental results very well and the predicted values are quite accurate by taking only 5 min or less. The satisfying results suggest that the introduction of concentration polarization theory might provide researchers an alternative model form to establish voltage relaxation models.

Abstract: The open-circuit voltage (OCV) of a battery, as a crucial characteristic parameter, is widely used in many aspects of battery technology, such as electrode material mechanism analysis, battery performance/state estimation and working process management. However, the applications of OCV are severely limited due to the need for a long rest time for full relaxation. In this paper, a rapid OCV prediction method is proposed to predict the final static OCV in a few minutes using linear regression techniques, based on a new mathematical model developed from an improvement on a second-order resistance–capacitance (RC) model. As the improvement, an important discovery is demonstrated by experimental investigation and data analysis: the relaxation time (i.e., time constant) of the diffusion circuit of the second-order RC model is not a fixed constant, unlike an intrinsic value for a given material, but an apparent linear function of the open-circuit time. This improvement enables the new model to track the actual relaxation process very well. The accuracy and the rapidity of the new model and proposed method are validated with working-condition experimental data on battery cells with different cathodes, and the results of OCV prediction are very accurate (errors below 1 mV in 20 min). [Copyright &y& Elsevier]

• To develop a reliable linear model for rapid prediction of temperature fields. • A dimensionless heat exchange ratio 〈 β 〉 proposed for decoupling of momentum and energy equations when increasing to infinity. • Linear temperature model (LTM) constructed and validated with lower ranges of 〈 β 〉 (compared to infinity). • To facilitate at fast and efficient predictions of temperature and HVAC control. The prediction of non-uniform thermal environment plays an important role in the control of heating, ventilation and air conditioning (HVAC) system, to meet the growing demand for occupant thermal comfort and building energy consumption, in response to varying heat sources. The popular approach to the prediction issue is computational fluid dynamics (CFD), with considerable time and computational costs. Thus, this work aims to develop a reliable linear model for rapid prediction of temperature fields, i.e., linear temperature model (LTM). A dimensionless heat exchange ratio 〈 β 〉 (when increasing to infinity) is proposed for the decoupling of momentum and energy equations, which is defined as the heat emission of the heat source to the heat gain of the flow. This will further be used for the construction of LTM combining Green's function and energy equation. By using the simulation method, a ventilation case occupied with various heat sources is considered for the validation, i.e., lower ranges of 〈 β 〉 (compared to infinity) as well as the related prediction error of LTM for engineering application. It is found that 〈 β 〉 can be valid with the value above 2.7 when source intensity is equal to 34 W/m3. Using the curve fitting, the conjunction between 〈 β 〉 , source intensity and prediction error is suggested for the verification of feasible LTM. The LTM can be applicative to a full-scale environment, to improve physical environment by 68% and energy efficiency by 50%. This work will facilitate at fast and efficient predictions of temperature distribution and HVAC systems online control. [ABSTRACT FROM AUTHOR]

The open circuit voltage of lithium ion batteries in equilibrium state, as a vital thermodynamic characteristic parameter, is extensively studied for battery state estimation and management. However, the time-consuming relaxation process, usually for several hours or more, seriously hinders the widespread application of open circuit voltage. In this paper, a novel voltage relaxation model is proposed to predict the final open circuit voltage when the lithium ion batteries are in equilibrium state with a small amount of sample data in the first few minutes, based on the concentration polarization theory. The Nernst equation is introduced to describe the evolution of relaxation voltage. The accuracy and effectiveness of the model are verified using experimental data on lithium ion batteries with different kinds of electrodes (LiCoO2/mesocarbon-microbead and LiFePO4/graphite) under different working conditions. The validation results show that the presented model can fit the experimental results very well and the predicted values are quite accurate by taking only 5 min or less. The satisfying results suggest that the introduction of concentration polarization theory might provide researchers an alternative model form to establish voltage relaxation models. [ABSTRACT FROM AUTHOR]

In this paper, early mouldy grain rapid prediction method using probabilistic neural network (PNN) and electronic nose (e-nose) was studied. E-nose responses to rice, red bean, and oat samples with different qualities were measured and recorded. E-nose data was analyzed using principal component analysis (PCA), back propagation (BP) network, and PNN, respectively. Results indicated that PCA and BP network could not clearly discriminate grain samples with different mouldy status and showed poor predicting accuracy. PNN showed satisfying discriminating abilities to grain samples with an accuracy of 93.75%. E-nose combined with PNN is effective for early mouldy grain prediction.